11

.. Chapter d t 5
The power series L::'=o anx
n
converges absolutely within the interval of
convergence Ixl < R, where
I
an I R = lim -- .
n-too a n+l
Jr'
COMPLEX NUMBERS
Relations between functions will be satisfied order by order when they are
replaced by their power series. You must know how to expand functions of
functions , out to some desired order within the common interval of conver-
gence.
jlr.
The power series representing a function may be integrated term by term and
differentiated term by term within the interval of convergence to obtain the
series for the integral or derivative of the function in question.
;jii.
ili Pr!
,lh
5.1. Introduction
Let us consider the quadratic function f {x) = x
2
- 5x+6 and ask where it vanishes .
If we plot it against x , we will find that it vanishes at x = 2 and x = 3. This is
also clear if we write f in factorized form as f{ x) = (x - 2){x - 3) . We could
equivalently use the well-known formula for the roots x of a quadratic equation:
89
A plot will show that this function is always positive and does not vanish for any
point on the z -axis. We are then led to conclude that this quadratic equation has
no roots. Let us pass from the graphical procedure which gives no solution to the
algebraic one which does give some form of answer even now. It says
The problemof course is that we do not know what to make of H since there is
no real number whose square is Thus if we take the stand that a number is not
a number unless it is a real number, we will have to conclude that some quadratic
equations have roots and other do not.
.

n

;..J -
i!

o
(])
00",



,...l >,
r--.Q:i
(]) B

,...., . -

QQ
UU
>0
='"


g$
o"ffi
'EQ



-'25
.5 2

1l;;





:/J =

u
e."
(5.1.4)
(5.1.3)
(5.1.2)
(5.1.1)
x = - l H
2
x
2
+ x + 1 = O.
-b ,jb'1 - 4ac
x =
2a
to find the roots x = 2,3. Suppose instead we consider
ax
2
+ bx +c = 0
namely

:>;'t f.
:si)'
,}


f)I"
' .

90 Chapter 5 Complex Numbers 91
and refer to x and v as its real and imaginary parts and denote them by the symbols
Re z and 1m z: A number withjust Y '#0 is called a pure imaginary number. The
solution to our quadratic equation has a real part x = -! and an imaginary part
'J
:l.J
'J)

h



8
H

-
ti
(1)

:5;:
"8
"'- >-@
f....4 .... ::i
(1)
r"'I __

""
uu
>-0
C"'J

:"3
=Ol
gf;i
c-'-
o
'EQ
gg
a
.g8


;s.Q
'g'5

-
(5.2.11)
(5.2.8)
(5.2.9)
(5.2.10)
(5.2.4)
(5.2.5)
rule (5.2.6)
multiplication rule (5.2.7)
Z2 = x 2 +iY2 implies
X2
Y2
Z = x - iy
Xl +iYl
X2 +iY2, we define.
(Xl +X2) +i(Yl +Y2) addition
(X1X2 - Y1Y2) +i(X1Y2 +X2Ytl
Zl = Xl +iYl
Xl
Yl
Zl
Z2
Zl +Z2
ZlZ2
Suppose this were not true. This would imply Xl - X2 = i(Y2 - Yl). without
both of them vanishing separately. Squaring both sides. we would find a positive
definite left-hand side and a negative definite right-hand side. The only way to
avoid a contradiction is for both sides to vanish, giving us 0 = -0, which is
something we can live with. .
Now, given any real number z , we can associate with it a unique number -x,
called its negative. We can do that with a complex number Z = x + iy too, by
negating x and y. This number is called - z , But now we have an intermediate
choice in which we negate just Y: the result
pronounced "z-star" is called the complex conjugate of z, Some people like to
write it z and call it "z-bar".
Problem 5.2.1. VerifY that this is so.
It was emphasized that we must think of Z = x +iy as a single number. However
it is a single number which has two parts, which can be uniquely identified. Thus
although 7 = 5 +2 is a single number, the decomposition of 7 into 2 and 5 is not
unique. On the other hand z = 3 +4i has a real part 3 and an imaginary part 4
and we cannot move things back and forth between the real and imaginary parts
keeping the number fixed. Thus if two complex numbers are equal, their real parts
and imaginary parts are separately equal:
where in the last equation we have opened out the brackets as we do with real
numbers and used i
2
= -1. Ifwe use these rules, we can verify that x =
indeed satisfies Eqn. (5.U). Note that ZlZ2 = Z2Zl.
y = V3/2. We think of z +iy as a single number. Indeed, if the number inside
the radical had been 3 instead of -3, surely we would have treated say - +V3/2
as a single number. as one of the roots. The same goes for +iV3/2.
The rules obeyed by complex numbers are as follows. Given two of them Zl
and Z2 .
(5.2.3)
(5.2.2)
(5.2.1) i = P .
Z = x +iy
IV3
x = - 2' T
i
.
We will postulate that i will behave like a real number in all manipulations involving
addition and multiplication and that the only new feature it will have is the one
that defined it, namely that its square equals - 1.
We now introduce a general complex number
In terms of i, the answer to Eqn. (5.1.3) is
This is how it was for many centuries until the rather bold suggestions was made
that we admit into the fold of numbers also those of the fonn R. All we need to
know to do this is a set of consistent rules for manipulating such numbers; being
able to visualize them is no prerequisite. It is really up to us make up the rules
since these entities have come out of the blue. The rules must. however. be free of
contradictions. Of course. all this is pointless if the whole enterprise does nothing
but merely exist. In the present case the idea has proven to be a very seminal one
and we will see some of the evidence even in this elementary treatment.
We are dealing here with a case of mathematical abstraction or generalization.
an example ofwhich you have already seen. when we extended the notion of powers
from integers to all real values. and examples of which you will see more than
once in this course. say when we extend the notion of vectors in three dimensions
with real components (which we can readily imagine) to vectors in any number
of dimensions. or vectors with complex components (like H) which we cannot
visualize. A general guideline when embarking on such generalizations is that we
impose on the new entities as many properties of the more familiar entities as is
possible. For example. when we passed from integer powers an to arbitrary powers
a
Z
(whatever that meant) we demanded that noninteger powers obey the same rule
of composition, i.e., aZa
ll
= a
Z
+
1I
for all x and y.
Returning to our problem we will first demand that square roots of negative
numbers (whatever they mean) still obey the rule that ..rc;b = .,faVb. Thus R =
V3R. The point of this is that the problem of taking the square root of any
negative number is reduced to taking the root of -1. Thus the basic building block
we need to introduce. called the unit imaginary number, is
5.2. Complex Numbers in Cartesian Form
92 Chapter 5 Complex Numbers . 93
This result is more easily obtained by using the notion of complex conjugates .
First note that
as you may check by explicit evaluation of both sides. The same obviously is
true for the sum of two complex numbers. Now, if two complex numbers are
Note that we did not ever explicitly define the rules for division of complex
numbers. This is because we can carry out division as the inverse of multiplication.
Thus if ZI / Z2 = Z3, we can multiply both sides by Z2 (which we know how to do)
and solve for X3 and Y3 by equating real and imaginary parts in
(5.2.28)
(5.2.27)
(5.2.26)
(5.2.25)
(5.2.24)
(5.2.23)
(5.2.22)
(5.2.21)
(V2-1)+ i(V2+1)
V2(1 +i)
[(V2 - 1) + i (V2 + 1)][1- i]
2V2
i
1 + V2'
(V2-1)+i (V2 +1) l+i
2 . V2
(V2 -1) + i (V2 + 1) + i(V2 -1) + i
2(
V2 + 1)
2V2
Z2
z l
1
i - V2
(V2 + i )( 1 + i) (V2 - 1) + i( V2 + 1)
Zl = (1 _ i )( 1 +i ) = 2
ZlZ2 =
Problem 5.2.3. Solve for x and y given
2 + 3i 2 2 9'
- - + - - = + s,
6 + 7i x +iy
o

='"
&.s

o Q
=ll

0"'3
II

g.Q

H
>'-

h
::; ...J

i2=
0=
58


-


EQ
Problem 5.2.2. Show that = i . - ;
=ll
. The following exercises should give you some more practice with the manip- (l)
ulation of complex numbers. '5i;;;
!i-

=0
"",
o

O

>.-
88
and
Since Z2 is already in this form, let us move on to compute
V2+ i l +i
Zl = -- Z2 = - -
I - i V2
Let us first simplify Zl to the form x + i y by multipl ying the numerator and
denominator by the complex conjugate of the latter:
equal, their real imaginary parts are separately equal. Consider an equation with
complex numbers on both sides. Ifwe repl ace all the numbers by their conjugates,
the resulting quantities must still be equal. This is because the imaginary parts,
originally equal on both sides, will continue to be equal after signs are changed on
both sides. Thus every complex equation implies another one obtained by complex
conjugation of both sides. The latter does not contain any more or less information:
both tell us the real and imaginary parts are separately equal.
Let us consider an illustrative example. Let
(5.2.20)
(5.2.17)
(5.2.16)
(5.2.15)
(ZIZ2) * =Ziz2
1 z *
- =
z zz*
(5.2.18)
x - i y
x 2 +y
2
'
(5.2.19)
Y3
X3
Re z == x
z + z"
(5.2.13) =
2
1m z == Y
z - z"
(5.2.14) =
2i
Xl + iYI = ( X3 + i Y3 )(X2 + i Y2 )
Note that
Applying this result to the I/Z2 in the ratio Zl/z2, we can obtain Eqns. (5.2.16,
5.2.17) more easily than before. In other words, by invoking the complex conjugate
we have reduced the problem to division by a real number, namely z *z , which is
a familiar concept.
Complex conjugation can be viewed as the process ofreplacing i by - i within
the complex number. Stated this way it is clear that the complex conjugate of a
product is the product of the complex conjugates
to obtain (upon solving a pair of simultaneous equations)
Xl X2 +Y IY2

Yl X2 - Y2 XI

zz" =. x
2
+ y
2
== Izl
2
O. (5.2.12)
One refers to Izi = ";x
2
+y2 as the modulus or absolute value of the complex
number z . It is useful to know that given z and its z *, we can recover the real and
imaginary parts of z as follows:
94
Chapter 5 Complex Numbers 95
Problem 5.2.5. Show that a polynomial with real coefficients has only real roots
or complex roots that come in complex conjugate pairs.
Problem 5.2.4. Find the real part, imaginary part. modulus. complex conjugate,
and inverse ofthefollowing numbers: (i) 3;4i' (ii) (3+4i)2. (iii) (iv)
and (v) cosO + isinO.
Problem 5.2.6. (Very Important). Prove algebraically that
IRe zl Izi
11m zl Izi .
(5.2.29)
(5.2.30)
y
z =x+ly = r e 10
i(' I'" x
z*=x -Iy
' IZl + z21
2
= IZll2 + IZ212 + 2 Re(zlz2) (5.2.31)
Figure 5.1. The complex plane,
5.3. Polar Form of Complex Numbers
where we will choose 0 to be real. To prove this identity, we must define what
we mean by e raised to a complex power iO. We define eanything to be the
infinite power series associated with the exponential function eX with x replaced
by anything. Thus
eelephant = (el ephant)n (5.3.2)
L..J n!
n=O
Problem 5.2.7. (Important). VerifY that the numbers Zl. Z2 from Eqn. (5.2.21)
respect Eqns. (5.2.31-5.2.33).
Recall that all real numbers can be visualized as points on a line, called the x axis.
Tovisualizeall complex numbers we introduce the complex plane which is just the
x - y plane. The complex number z = x + iy is labeled as shown in Fig. 5.1. The
conjugate is z*. The significance of rand 0 will now be explained.

ES

o
=;a
50

g 51

Yo



g.


>!-

h
1-;


::.J :::
58


.:;;
CO
"'''

(l)
:Ii",

:5;:
08

>-;;
""toi
(l)

O

88
(5.3.7)
(5.3.5)
(5.3.4)
(5.3.6)
(5.3.3)
e
i 7r
+ 1 = O.
z = x +iy
Who would have thought that 1r which enters as the ratio of circumference to
diameter, e, as the natural base for logarithms, i, as the fundamental imaginary unit
and 0 and 1 (which we know all about from infancy) would all be tied together in
any way, not to mention such a simple and compact way? I hope I never stumble
into anything like this formula, for nothing I do after that in life would have any
significance.
Look at Fig. 5.1 of the complex plane and note that
where we have used the fact that i
2
= -I, i
3
= - i , i
4
= I, and so on, as well
as the infinite series that define the sine and cosine functions. (These expansions
converge for all finite IJ, as shown before. The presence of i does not in any way
complicate the question of convergence since it either turns into a I or into i .)
Setting 0 = 1r we obtain one of the most remarkable formulae inmathematics:
which converges for any finite sized el ephant.
Turning to our problem, we expand the infinite series for the exponential and
collect the real and imaginary parts as follows:
e
i8
= f
n=O n .
00 (_1)n(IJ)2n . 00 (_I)n(IJ)2n+1
= L (2n)! +tL '- ..
n=O n=O
= coslJ + i sinlJ

(5.3.1)
(5.2.32)
(5.2.33) !ZlZ21 = IZlll z21
e
i 9
= cos s + isin O,
IZl + z21 Iztl + IZ21
We begin this section with a remarkable identity due to Euler:
96 Chapter5 Complex Numbers
97
--Io multiply two complex numbers, multiply their moduli and add their phases.
To divide. divide by the modulus and subtract the phase of the denominator.
z
=
re
i 9
then (5.3.11)
z
=
re-
i9
(5.3.12)
zz
=
r
2
(Izl =r) (5.3.13)
1
1 -i9
(5.3.14) -
=
-e
z r
ZlZ2
=
rlr2e i (91+ 9. ) .
(5.3.15)
(We are using the fact that the law of composition of exponents under a product
works for complex exponents as well. Indeed this is built into the exponential
function defined by the infinite series. You may check that this works to any
given order even for imaginary arguments. In Chapter 6, this will be proven more
directly.) The last formula tells us how easy it is to multiply or divide two complex
numbers in polar form:
The last equation is called the polar form of the complex number as compared to
the cartesian form we have been using so far. One refers to e as the argument or
phase of the number and r as its modulus or absolute value. It is just as easy to
visualize the number in the complex plane given the polar form as it was with the
cartesian form. (As is true with polar coordinates in any context, 0 is defined only
modulo 211", that is to say adding 21r to it changes nothing. We can usually restrict
it to the interval [0- 211"] .) All manipulations we did before in the cartesian form
can of course be carried out in polar form, though some become easier and some
harder. Thus if
>0


e
=;a
= fj,
0,



g
u 0
u
:;5


-gil

1/-

'ji
U



.


-
i!

o
(]) :;
OO"/l



,....l '-iil
r--.Q:i
(]) B

,...., . -

00
uu
(5.3.24)
(5.3.25)
(5.3.26)
(5.3.27)
(5.3.22)
(5.3.21)
(5.3.20)
(5.3.19)
o (1)
11" / 2 (i)
11" (-1)
-1r/2 (- i ).
o
1
Problem 5.3.2. For the fol/owing pairs of numbers. give their polar form. their
complex conjugates. their moduli. product. the quotient zI! Z2 and the complex
conjugate of the quotient:
You are expected to know these points at all times.
Problem 5.3.1. VerifY the correctness of the above using Euler 's f ormula.
When we work with real numbers, we know that multiplication by a number, say
4, rescales the given number by 4. Multiplying a number in the complex plane by
r ei8. rescales its length (or modulus) by r and also rotates it counterclockwise by
0. Multiplying by a unimodular number simply rotates without any rescaling.
l+i .
Zl = v'2 Z2 = V3 - I
Zl = 3 + 4i Z2 = [1 + 2i ]-2
3 - 4i 1 - 3i
- Z2 = + =e
i 1f
/
4
=e
785i
.
Now it is easy to form the product and quotient
ZlZ2 = l e (1.400+.785)i
= 1.224 (cos 2.185 + i sin 2.185)
1 .
--+1
v'2
f!e (1.400-.785)i = 1.224 (cos .615 + i sin .615)
V"2 .
i
1+ v'2 (5.3.23)
in agreement with the calculation done earlier in cartesian form.
Complex numbers Z =r e
i 8
with r =1 have Izi =1 and are called unimodular.
We may imagine them as lying on a circle of unit radius in the complex plane.
Special points on this circle are
As for Z2,

(5.3.8)
(5.3.9)
(5.3.10)

r [cos0+ i sin 0)
-.
On the other hand to add two complex numbers we have to go back to the
cartesian form, add the components and revert to the polar form.
Let us return to Eqn, (5.2.21) and manipulate the numbers in polar form. First
Zl = (5.3.16)
(v'2-1)2+(v'2+1)2 [_ [1+v'2]] (5317)
4
exp 1 arctan . In .
y2-1
(5.3.18)

_.
--
,.

."-.-"
..., - 2;
," -001' +'''.',
,".. . -1,--".
."+.-"
""' , - 2
r ...... tII ....... y...... -....
The """"I<:< """IiuPI< it
0... call ond, .... phu< or 0lJ'l'l"'I' of .
EoI.. by lUina,..l ..... imr.J;ioary pons. > fiDd
(lJ.28)
.. . ;0 , " ... . - 10
.. ""' . .. 2
.''_._If
_ 1"' ... .. 2; .
, " 2<;' / ' <, _ 6<" />.

C'Jooct./OI-.J-Ilw,""
..,,' . +ooti' . _I. .... 2. _201. ..... . .
_2<_....'._.',.
Tho modiulw >qIaRIIl 1I Ii- by
1-1' -,' +,' .. ..-
_ _ _ oM

o.d+ _ 31 +....;. <00(,., - 1)1 .. -
201
""" -. . 1/lIJ//dT _ "*'" _ Mdt ." ......' df"" .

I', +" I' - 1.,1' +1_,1' + U\e\.,.;)
I>, +..1S 1.,1 +1.,1 1., ,,1-1',11,,1
(. ,.,)' .. ' i';
.. 2 . ..,-.. ).
" "
s..s. Summary
. "
....... ........... port'" Roo ' " II .... imq:i.., pori Of 1m . Thio II ....
canesian ""'" of .
The pOI.- _ ..
. .. j;f+';J, 1&11 , .../_,